[AMDGPU] TTI: Prioritize per-iter base adds in LSR plan comparison

.azuredevops/rocm-ci.yml

@@ -0,0 +1,29 @@
+resources:
+  repositories:
+  - repository: pipelines_repo
+    type: github
+    endpoint: ROCm
+    name: ROCm/ROCm
+
+variables:
+- group: common
+- template: /.azuredevops/variables-global.yml@pipelines_repo
+
+trigger:
+  batch: true
+  branches:
+    include:
+    - amd-staging
+    - amd-mainline-open
+  paths:
+    exclude:
+    - .ci
+    - .github
+    - '*.md'
+    - '**/*.md'
+    - LICENSE.TXT
+
+pr: none
+
+jobs:
+  - template: ${{ variables.CI_COMPONENT_PATH }}/llvm-project.yml@pipelines_repo

.ci/generate_test_report_github.py

@@ -8,6 +8,7 @@
 
 import generate_test_report_lib
 
+
 def compute_platform_title() -> str:
     logo = ":window:" if platform.system() == "Windows" else ":penguin:"
     # On Linux the machine value is x86_64 on Windows it is AMD64.

.ci/generate_test_report_lib.py

@@ -41,10 +41,12 @@ def _parse_ninja_log(ninja_log: list[str]) -> list[tuple[str, str]]:
         # touch test/4.stamp
         #
         # index will point to the line that starts with Failed:. The progress
-        # indicator is the line before this ([4/5] test/4.stamp) and contains a pretty
-        # printed version of the target being built (test/4.stamp). We use this line
-        # and remove the progress information to get a succinct name for the target.
-        failing_action = ninja_log[index - 1].split("] ")[1]
+        # indicator is sometimes the line before this ([4/5] test/4.stamp) and
+        # will contain a pretty printed version of the target being built
+        # (test/4.stamp) when accurate. We instead parse the failed line rather
+        # than the progress indicator as the progress indicator may not be
+        # aligned with the failure.
+        failing_action = ninja_log[index].split("FAILED: ")[1]
         failure_log = []
         while (
             index < len(ninja_log)
@@ -98,6 +100,7 @@ def _format_ninja_failures(ninja_failures: list[tuple[str, str]]) -> list[str]:
         )
     return output
 
+
 def get_failures(junit_objects) -> dict[str, list[tuple[str, str]]]:
     failures = {}
     for results in junit_objects:

.ci/generate_test_report_lib_test.py

@@ -39,7 +39,7 @@ def test_find_failure_ninja_logs(self):
         self.assertEqual(
             failures[0],
             (
-                "test/4.stamp",
+                "touch test/4.stamp",
                 dedent(
                     """\
                     FAILED: touch test/4.stamp
@@ -77,7 +77,7 @@ def test_ninja_log_end(self):
         self.assertEqual(
             failures[0],
             (
-                "test/3.stamp",
+                "touch test/3.stamp",
                 dedent(
                     """\
                     FAILED: touch test/3.stamp
@@ -106,7 +106,7 @@ def test_ninja_log_multiple_failures(self):
         self.assertEqual(
             failures[0],
             (
-                "test/2.stamp",
+                "touch test/2.stamp",
                 dedent(
                     """\
                     FAILED: touch test/2.stamp
@@ -117,7 +117,7 @@ def test_ninja_log_multiple_failures(self):
         self.assertEqual(
             failures[1],
             (
-                "test/4.stamp",
+                "touch test/4.stamp",
                 dedent(
                     """\
                     FAILED: touch test/4.stamp
@@ -150,7 +150,7 @@ def test_ninja_log_runtimes_failure(self):
         self.assertEqual(
             failures[0],
             (
-                "test/2.stamp",
+                "touch test/2.stamp",
                 dedent(
                     """\
                     FAILED: touch test/2.stamp
@@ -159,6 +159,34 @@ def test_ninja_log_runtimes_failure(self):
             ),
         )
 
+    # Test that we correctly handle cases where the FAILED: line does not
+    # match up with the progress indicator.
+    def test_ninja_log_mismatched_failed(self):
+        failures = generate_test_report_lib.find_failure_in_ninja_logs(
+            [
+                [
+                    "[1/5] test/1.stamp",
+                    "[2/5] test/2.stamp",
+                    "ModuleNotFoundError: No module named 'mount_langley'",
+                    "FAILED: tools/check-langley",
+                    "Wow! This system is really broken!",
+                    "[5/5] test/5.stamp",
+                ]
+            ]
+        )
+        self.assertEqual(len(failures), 1)
+        self.assertEqual(
+            failures[0],
+            (
+                "tools/check-langley",
+                dedent(
+                    """\
+                    FAILED: tools/check-langley
+                    Wow! This system is really broken!"""
+                ),
+            ),
+        )
+
     def test_title_only(self):
         self.assertEqual(
             generate_test_report_lib.generate_report("Foo", 0, [], []),
@@ -407,7 +435,6 @@ def test_no_failures_multiple_build_failed_ninja_log(self):
                 ]
             ],
         )
-        print(test)
         self.assertEqual(
             generate_test_report_lib.generate_report(
                 "Foo",
@@ -449,15 +476,15 @@ def test_no_failures_multiple_build_failed_ninja_log(self):
                     All tests passed but another part of the build **failed**. Click on a failure below to see the details.
 
                     <details>
-                    <summary>test/2.stamp</summary>
+                    <summary>touch test/2.stamp</summary>
 
                     ```
                     FAILED: touch test/2.stamp
                     Wow! Be Kind!
                     ```
                     </details>
                     <details>
-                    <summary>test/4.stamp</summary>
+                    <summary>touch test/4.stamp</summary>
 
                     ```
                     FAILED: touch test/4.stamp

.ci/monolithic-windows.sh

@@ -23,8 +23,8 @@ runtimes_targets="${4}"
 start-group "CMake"
 pip install -q -r "${MONOREPO_ROOT}"/.ci/all_requirements.txt
 
-export CC=cl
-export CXX=cl
+export CC=C:/clang/clang-msvc/bin/clang-cl.exe
+export CXX=C:/clang/clang-msvc/bin/clang-cl.exe
 export LD=link
 
 # The CMAKE_*_LINKER_FLAGS to disable the manifest come from research
@@ -49,6 +49,7 @@ cmake -S "${MONOREPO_ROOT}"/llvm -B "${BUILD_DIR}" \
       -D CMAKE_EXE_LINKER_FLAGS="/MANIFEST:NO" \
       -D CMAKE_MODULE_LINKER_FLAGS="/MANIFEST:NO" \
       -D CMAKE_SHARED_LINKER_FLAGS="/MANIFEST:NO" \
+      -D CMAKE_CXX_FLAGS="-Wno-c++98-compat -Wno-c++14-compat -Wno-unsafe-buffer-usage -Wno-old-style-cast" \
       -D LLVM_ENABLE_RUNTIMES="${runtimes}"
 
 start-group "ninja"

.ci/premerge_advisor_explain.py

@@ -40,7 +40,9 @@ def main(commit_sha: str, build_log_files: list[str]):
             explanation_request["failures"].append(
                 {"name": name, "message": failure_message}
             )
-    advisor_response = requests.get(PREMERGE_ADVISOR_URL, json=explanation_request)
+    advisor_response = requests.get(
+        PREMERGE_ADVISOR_URL, json=explanation_request, timeout=5
+    )
     if advisor_response.status_code == 200:
         print(advisor_response.json())
     else:

.ci/premerge_advisor_upload.py

@@ -12,9 +12,12 @@
 
 import generate_test_report_lib
 
-PREMERGE_ADVISOR_URL = (
-    "http://premerge-advisor.premerge-advisor.svc.cluster.local:5000/upload"
-)
+# These are IP addresses of the two premerge advisor instances. They should
+# eventually be updated to domain names.
+PREMERGE_ADVISOR_URLS = [
+    "http://34.82.126.63:5000/upload",
+    "http://136.114.125.23:5000/upload",
+]
 
 
 def main(commit_sha, workflow_run_number, build_log_files):
@@ -41,7 +44,8 @@ def main(commit_sha, workflow_run_number, build_log_files):
         ninja_failures = generate_test_report_lib.find_failure_in_ninja_logs(ninja_logs)
         for name, failure_message in ninja_failures:
             failure_info["failures"].append({"name": name, "message": failure_message})
-    requests.post(PREMERGE_ADVISOR_URL, json=failure_info)
+    for premerge_advisor_url in PREMERGE_ADVISOR_URLS:
+        requests.post(premerge_advisor_url, json=failure_info, timeout=5)
 
 
 if __name__ == "__main__":

.ci/utils.sh

@@ -40,13 +40,17 @@ function at-exit {
   fi
 
   if [[ "$retcode" != "0" ]]; then
-    python "${MONOREPO_ROOT}"/.ci/premerge_advisor_upload.py \
-      $(git rev-parse HEAD~1) $GITHUB_RUN_NUMBER \
-      "${BUILD_DIR}"/test-results.*.xml "${MONOREPO_ROOT}"/ninja*.log
     if [[ "$GITHUB_ACTIONS" != "" ]]; then
       python "${MONOREPO_ROOT}"/.ci/premerge_advisor_explain.py \
         $(git rev-parse HEAD~1) "${BUILD_DIR}"/test-results.*.xml \
         "${MONOREPO_ROOT}"/ninja*.log
+      python "${MONOREPO_ROOT}"/.ci/premerge_advisor_upload.py \
+        $(git rev-parse HEAD~1) $GITHUB_RUN_NUMBER \
+        "${BUILD_DIR}"/test-results.*.xml "${MONOREPO_ROOT}"/ninja*.log
+    else
+      python "${MONOREPO_ROOT}"/.ci/premerge_advisor_upload.py \
+        $(git rev-parse HEAD) $BUILDBOT_BUILDNUMBER \
+        "${BUILD_DIR}"/test-results.*.xml "${MONOREPO_ROOT}"/ninja*.log
     fi
   fi
 }

.github/CODEOWNERS

@@ -168,3 +168,8 @@
 
 # libclang/Python bindings
 /clang/bindings/python @DeinAlptraum
+
+# AMD-specific projects.
+/amd/comgr/ @jlambert_amdeng
+/amd/device-libs/ @bsumner_amdeng
+/amd/hipcc/ @dsalinas_amdeng

.github/PULL_REQUEST_TEMPLATE.md

@@ -0,0 +1 @@
+

.github/dependabot.yml

@@ -0,0 +1,27 @@
+version: 2
+updates:
+  - package-ecosystem: "github-actions"
+    directory: "/"
+    schedule:
+      interval: "monthly"
+    groups:
+      github-actions:
+        patterns:
+          - "*"
+  - package-ecosystem: "pip"
+    directory: "/llvm/docs"
+    schedule:
+      interval: "monthly"
+    groups:
+      llvm-docs-requirements:
+        patterns:
+          - "*"
+  - package-ecosystem: "pip" # See documentation for possible values
+    directory: "/amd/hipcc/docs/sphinx" # Location of package manifests
+    open-pull-requests-limit: 10
+    schedule:
+      interval: "daily"
+    labels:
+      - "documentation"
+    reviewers:
+      - "samjwu"

.github/instructions/lldb.instructions.md

@@ -0,0 +1,79 @@
+---
+applyTo: lldb/**/*
+---
+
+When reviewing code, focus on:
+
+## Language, Libraries & Standards
+
+- Target C++17 and avoid vendor-specific extensions.
+- For Python scripts, follow PEP 8.
+- Prefer standard library or LLVM support libraries instead of reinventing data structures.
+
+## Comments & Documentation
+
+- Each source file should include the standard LLVM file header.
+- Header files must have proper header guards.
+- Non-trivial classes and public methods should have Doxygen documentation.
+- Use `//` or `///` comments normally; avoid block comments unless necessary.
+- Non-trivial code should have comments explaining what it does and why. Avoid comments that explain how it does it at a micro level.
+
+## Language & Compiler Issues
+
+- Write portable code; wrap non-portable code in interfaces.
+- Do not use RTTI or exceptions.
+- Prefer C++-style casts over C-style casts.
+- Do not use static constructors.
+- Use `class` or `struct` consistently; `struct` only for all-public data.
+- When then same class is declared or defined multiple times, make sure it's consistently done using either `class` or `struct`.
+
+## Headers & Library Layering
+
+- Include order: module header → local/private headers → project headers → system headers.
+- Headers must compile standalone (include all dependencies).
+- Maintain proper library layering; avoid circular dependencies.
+- Include minimally; use forward declarations where possible.
+- Keep internal headers private to modules.
+- Use full namespace qualifiers for out-of-line definitions.
+
+## Control Flow & Structure
+
+- Prefer early exits over deep nesting.
+- Do not use `else` after `return`, `continue`, `break`, or `goto`.
+- Encapsulate loops that compute predicates into helper functions.
+
+## Naming
+
+- LLDB's code style differs from LLVM's coding style.
+- Variables are `snake_case`.
+- Functions and methods are `UpperCamelCase`.
+- Static, global and member variables have `s_`, `g_` and `m_` prefixes respectively.
+
+## General Guidelines
+
+- Use `assert` liberally; prefer `llvm_unreachable` for unreachable states.
+- Do not use `using namespace std;` in headers.
+- Provide a virtual method anchor for classes defined in headers.
+- Do not use default labels in fully covered switches over enumerations.
+- Use range-based for loops wherever possible.
+- Capture `end()` outside loops if not using range-based iteration.
+- Including `<iostream>` is forbidded. Use LLVM’s `raw_ostream` instead.
+- Don’t use `inline` when defining a function in a class definition.
+
+## Microscopic Details
+
+- Preserve existing style in modified code.
+- Prefer pre-increment (`++i`) when value is unused.
+- Use `private`, `protected`, or `public` keyword as appropriate to restrict class member visibility.
+- Omit braces for single-statement `if`, `else`, `while`, `for` unless needed.
+
+## Review Style
+
+- Be specific and actionable in feedback.
+- Explain the "why" behind recommendations.
+- Link back to the LLVM Coding Standards: https://llvm.org/docs/CodingStandards.html.
+- Ask clarifying questions when code intent is unclear.
+
+Ignore formatting and assume that's handled by external tools like `clang-format` and `black`.
+Remember that these standards are **guidelines**.
+Always prioritize consistency with the style that is already being used by the surrounding code.

.github/instructions/llvm.instructions.md

@@ -0,0 +1,8 @@
+---
+applyTo: llvm/**/*
+---
+
+When performing a code review, pay close attention to code modifying a function's
+control flow. Could the change result in the corruption of performance profile
+data? Could the change result in invalid debug information, in particular for
+branches and calls?